End-to-End Fine-Grained Action Segmentation and Recognition Using Conditional Random Field Models and Discriminative Sparse Coding.
Effrosyni MavroudiDivya BhaskaraShahin SefatiHaider AliRené VidalPublished in: CoRR (2018)
Keyphrases
- fine grained
- end to end
- sparse coding
- random field models
- sparse codes
- discriminative power
- object category recognition
- dictionary learning
- natural images
- coarse grained
- unsupervised learning
- random fields
- sparse representation
- image classification
- linear combination
- discriminative dictionary
- multiscale
- generative model
- image representation
- access control
- image patches
- segmentation algorithm
- segmentation method
- image segmentation
- autoregressive
- feature space
- visual words
- high dimensional
- object recognition
- bag of words
- test images
- conditional random fields
- co occurrence
- reinforcement learning
- feature extraction